• 한국정보통신학회논문지
• /
• 제14권7호
• /
• pp.1557-1564
• /
• 2010

감성신호는 개인에 따라 그 패턴이 매우 다르게 나타나므로, 본 논문에서는 감성신호의 개인별 특징을 고려한 최대 엔트로피 기반 감성 인식 모델을 제안한다. 제안하는 모델은 보다 정확하게 사용자의 감성을 인식하기 위해서, 단순히 주어진 입력 감성 신호 값만을 사용하지 않고, 긍정 감성 신호 값의 평균과 부정 감성 신호 값의 평균을 입력 감성 신호의 값과 비교하여 활용한다. 또한, 감성 인식에 대한 전문적인 지식이 없이도 감성 인식 모델의 구축이 용이하도록, 제안하는 모델은 성능이 높다고 잘 알려진 기계학습기법의 하나인 최대 엔트로피 모델을 이용한다. 감성 신호의 수치 값을 그대로 사용하면 기계 학습에 필요한 학습 패턴 자료를 충분히 확보하기 어렵다는 점을 고려하여, 제안하는 모델은 평균차를 수치 값 대신 +(양수)와 -(음수)로 단순하게 표현하며, 감성 반응 전체 시간인 10초 대신 초단위로 분할하여 학습 패턴 자료의 양을 늘렸다.

• 대기
• /
• 제25권3호
• /
• pp.511-520
• /
• 2015

Existing studies on radar rainfall uncertainties were performed to reduce the uncertainty for each stage by using bias correction during the quantitative radar rainfall estimation process. However, the studies do not provide quantitative comparison with the uncertainties for all stages. Consequently, this study proposes a suitable approach that can quantify the uncertainties at each stage of the quantitative radar rainfall estimation process. First, the new approach can present initial and final uncertainties, increasing or decreasing the uncertainty, and the uncertainty percentage at each stage. Furthermore, Maximum Entropy (ME) was applied to quantify the uncertainty in the entire process. Second, for the uncertainty quantification of radar rainfall estimation at each stage, this study used two quality control algorithms, two rainfall estimation relations, and two bias correction techniques as post-processing and progressed through all stages of the radar rainfall estimation. For the proposed approach, the final uncertainty (ME = 3.81) from the ME of the bias correction stage was the smallest while the uncertainty of the rainfall estimation stage was higher because of the use of an unsuitable relation. Additionally, the ME of the quality control was at 4.28 (112.34%), while that of the rainfall estimation was at 4.53 (118.90%), and that of the bias correction at 3.81 (100%). However, this study also determined that selecting the appropriate method for each stage would gradually reduce the uncertainty at each stage. Finally, the uncertainty due to natural variability was 93.70% of the final uncertainty. Thus, the results indicate that this new approach can contribute significantly to the field of uncertainty estimation and help with estimating more accurate radar rainfall.

• 한국지능시스템학회논문지
• /
• 제18권4호
• /
• pp.501-506
• /
• 2008

통계적 방식의 품사부착 문제는 보통 N-그램과 같을 단일 통계정보를 활용하지만 단순한 통계 정보라는 원천적인 한계가 있어 많은 오류가 발생한다. 따라서 다양한 정보를 활용하는 것이 정확도를 높일 가능성이 있다는 데는 이론의 여지가 없다. 그러나 다른 종류의 통계 자료는 배타적 자료가 아닌 한 상충되는 정보를 가질 수밖에 없으므로 이러한 정보들로부터 어떻게 종합적인 결론을 내는가가 문제이다. 본 논문에서는 이러한 상이한 통계정보를 통합하는 방법으로 언어 모델의 구성에서 활용된 바 있는 최대엔트로피 모델의 한국어 품사 부착에의 사용 가능성을 제시한다. 여기서는 이종의 통계정보로서 N그램과, 트리거 쌍을 사용하게 된다. 이러한 트리거 쌍 통계정보를 N그램과 함께 최대엔트로피 모델링을 했을 경우 퍼플렉시티가 어떻게 변화하는지에 대한 실험결과를 관찰하게 될 것이다. 트리거 쌍은 또한 다양하게 문맥사이즈를 변화할 수 있으며, N그램의 확률 모델도 다양하기 때문에 여러 종류의 실험을 통한 많은 향상을 예상 할 수 있다. 본 실험에서는 단일 모델 사용시 94.9 %의 정확도를 가진 3-그램 모델에 트리거 쌍을 최대 엔트로피 방식으로 추가한 견우 95.6% 의 정확도를 보여 0.7% 포인트의 정확도 향상을 기록하였다. 따라서 향후 다양한 정보원을 개발하여 최대엔트로피 방식으로 통합할 경우 지속적인 정확도 향상을 가져 올 수 있는 가능성을 보여준다고 할 수 있다.

• Journal of Ecology and Environment
• /
• 제42권4호
• /
• pp.227-239
• /
• 2018

Background: Predicting the potential habitat of plants in arid regions, especially for medicinal ones, is very important. Although Pergularia tomentosa is a key species for medicinal purposes, it appears in very low density in the arid rangelands of Iran, needing an urgent ecological attention. In this study, we modeled and predicted the potential habitat of P. tomentosa using maximum entropy, and the effects of environmental factors (geology, geomorphology, altitude, and soil properties) on some characteristics of the species were determined. Results: The results showed that P. tomentosa was absent in igneous formation while it appeared in conglomerate formation. In addition, among geomorphological units, the best quantitative characteristics of P. tomentosa was belonged to the conglomerate formation-small hill area (plant aerial parts = 57.63 and root length = 30.68 cm) with the highest electrical conductivity, silt, and $CaCO_3$ content. Conversely, the species was not found in the mountainous area with igneous formation. Moreover, plant density, length of roots, and aerial parts of the species were negatively correlated with soil sand, while positive correlation was observed with $CaCO_3$, EC, potassium, and silt content. The maximum entropy was found to be a reliable method (ROC = 0.91) for predicting suitable habitats for P. tomentosa. Conclusion: These results suggest that in evaluating the plant's habitat suitability in arid regions, contrary to the importance of the topography, some environmental variables such as geomorphology and geology can play the main role in rangeland plants' habitat suitability.

• Current Applied Physics
• /
• 제18권12호
• /
• pp.1605-1608
• /
• 2018

$Gd_{1-x}Ho_xNi$ melt-spun ribbons were fabricated by a single-roller melt spinning method. All the compounds crystallize in an orthorhombic CrB-type structure. The Curie temperature ($T_C$) was tuned between 46 and 99 K by varying the concentration of Gd and Ho. A spin reorientation (SRO) transition is observed around 13 K. Different from $T_C$, the SRO transition temperature is almost invariable for all compounds. Two peaks of magnetic entropy change (${\Delta}S_M$) were found. One at the higher temperature range was originated from the paramagnet-ferromagnet phase transition and the other at the lower temperature range was caused by the SRO transition. The maximum of ${\Delta}S_M$ around $T_C$ is almost same. The other maximum of ${\Delta}S_M$ around SRO transition, however, had significantly positive relationship with x. It reached a maximum about $8.2J\;kg^{-1}\;K^{-1}$ for x = 0.8. Thus double large ${\Delta}S_M$ peaks were obtained in $Gd_{1-x}Ho_xNi$ melt-spun ribbons with the high Ho concentration. And the refrigerant capacity power reached a maximum of $622J\;kg^{-1}$ for x = 0.6. $Gd_{1-x}Ho_xNi$ ribbons could be good candidate for magnetic refrigerant working in the low temperature especially near the liquid nitrogen temperature range.

• Journal of Information Processing Systems
• /
• 제11권2호
• /
• pp.248-265
• /
• 2015

In this paper, we propose a maximum entropy-based model, which can mathematically explain the bio-molecular event extraction problem. The proposed model generates an event table, which can represent the relationship between an event trigger and its arguments. The complex sentences with distinctive event structures can be also represented by the event table. Previous approaches intuitively designed a pipeline system, which sequentially performs trigger detection and arguments recognition, and thus, did not clearly explain the relationship between identified triggers and arguments. On the other hand, the proposed model generates an event table that can represent triggers, their arguments, and their relationships. The desired events can be easily extracted from the event table. Experimental results show that the proposed model can cover 91.36% of events in the training dataset and that it can achieve a 50.44% recall in the test dataset by using the event table.

• Bulletin of the Korean Chemical Society
• /
• 제29권10호
• /
• pp.2017-2022
• /
• 2008

Of the experiments to shorten NMR measuring time by sparse sampling, non-uniform sampling (NUS) is advantageous. NUS miminizes systematic errors which arise due to the lack of samplings by randomization. In this study, I report the real-time acquisition of 3D NMR data using NUS and maximum-entropy (MaxEnt) data processing. The real-time acquisition combined with NUS can reduce NMR measuring time much more. Compared with multidimensional decomposition (MDD) method, which was originally suggested by Jaravine and Orekhov (JACS 2006, 13421-13426), MaxEnt is faster at least several times and more suitable for the realtime acquisition. The designed sampling schedule of current study makes all the spectra during acquisition have the comparable resulting resolutions by MaxEnt. Therefore, one can judge the quality of spectra easily by examining the intensities of peaks. I report two cases of 3D experiments as examples with the simulated subdataset from experimental data. In both cases, the spectra having good qualitie for data analysis could be obtained only with 3% of original data. Its corresponding NMR measuring time was 8 minutes for 3D HNCO of ubiquitin.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2007년도 Proceedings of ISRS 2007
• /
• pp.414-417
• /
• 2007

In this paper we present an adaptive method for accelerating conventional Maximum Entropy Method (MEM) for restoration of Passive Millimeter-Wave (PMMW) image from its blurred and noisy version. MEM is nonlinear and its convergence is very slow. We present a new method to accelerate the MEM by using an exponent on the correction ratio. In this method the exponent is computed adaptively in each iteration, using first-order derivatives of deblurred image in previous two iterations. Using this exponent the accelerated MEM emphasizes speed at the beginning stages and stability at later stages. In accelerated MEM the non-negativity is automatically ensured and also conservation of flux without additional computation. Simulation study shows that the accelerated MEM gives better results in terms of RMSE, SNR, moreover, it takes only about 46% lesser iterations than conventional MEM. This is also confirmed by applying this algorithm on actual PMMW image captured by 94 GHz mechanically scanned radiometer.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1990년도 춘계학술대회
• /
• pp.47-50
• /
• 1990

In this study, the degree of spectral transfer to lower frequency caused by accumulation of Latic acid inside the muscle is estimated the convintional dip analysis, zero-crossing method and FFT method have intrinsic errors and estimation problems in case of severe noise. The new spectral analysis method using "$2^{nd}$ order Maximum Entropy Method" was applied to estimate mean frequency and we confirmed that this new method yields fast and reliable estimation over the FFT method.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2010년도 춘계학술대회
• /
• pp.985-987
• /
• 2010

본 논문에서는 개인별 평균차를 이용한 최대 엔트로피 기반 감성 인식 모델을 제안한다. 정확하게 사용자의 감성을 인식할 수 있도록, 제안하는 모델은 단순하게 주어진 입력 감성 신호 정보만 분석하여 사용하지 않고, 입력 정보를 각 감성 상태의 평균값과 비교한 결과를 활용한다. 그리고, 자료 부족 문제를 완화하기 위해서, 제안하는 모델은 평균차를 +(양수)와 -(음수)로 단순하게 표현하고, 감성 반응 전체 시간 대신 초단위로 분할하여 감성신호의 평균을 계산한다. 또한, 전문적인 지식이 없이도 구축이 용이하도록, 제안하는 모델은 간단한 평균차 계산 기법과 잘 알려진 기계학습기법의 하나인 최대 엔트로피 모델을 이용한다.